Field
Embodiments of the present invention relate generally to a method, apparatus and system for rendering an information bearing function of time based on input signals. Embodiments of the present invention present a novel solution to increasing operational battery life and reduced thermal footprint of a cell phone or other communications device, apparatus, module, subsystem or component by using enhanced information distribution and power supply control techniques. More particularly, embodiments of the present invention are directed to rendering the information bearing function of time based on efficient information distribution and without requiring feedback loops or pre-distortion techniques.
Background Discussion
Currently, cell phones and other mobile communications platforms use an integral battery as a power source. The power source has limited storage capacity and users are increasingly demanding better performance from their cell phones. Generally, the cell phone transmitter, and particularly the power amplifier (PA), consumes a significant amount of battery power and generates the most heat when compared with other phone functions. The relative battery power demand of the PA is driven by the RF link budget and PA efficiency. The PA is not efficient because it transmits signals while operating in a substantially linear mode of operation. Both high power output and linearity are required to ensure that the transmissions from the cell phone conform to currently defined industry standards, and to overcome communication link budget deficits. Unwanted heat is generated by the PA because of inefficiencies in PA operation.
Generally, PAs operating in a linear mode, are not particularly efficient, and so, currently, a compromise must be made between battery life and conformance to the defined industry standards. Since the defined industry standards are mandatory and inflexible, the reduced battery life due to the PAs higher power consumption has been expected as a necessary impact.
Several conventional PA techniques have been developed in an attempt to improve operational efficiency of cell phones. Some examples of these conventional techniques include: envelope tracking; pre-distortion; feedback loops; and polar modulation. Other conventional approaches include amplification techniques, which include: Class AB power amplifiers; stage switching amplifiers, or Doherty amplifiers; envelope elimination and restoration amplifiers (EER); and outphasing and linear amplification with nonlinear components (LILAC) amplifiers. Each of these conventional techniques has drawbacks that make it inadequate. Thus, embodiments of the present invention have innovated in a different direction to overcome the inadequacies of the conventional approaches. Some conventional approaches are described below.
Envelope Tracking
An objective of envelope tracking is to improve the efficiency of power amplifiers (PA) carrying high peak-to average-power-ratio (PAPR) signals. The need to achieve high data throughput within limited spectrum resources requires the use of linear modulation with high peak to average power. Unfortunately, traditional fixed-supply power amplifiers operating under these conditions have low efficiency. One approach to improve the efficiency of a power amplifier is to vary the amplifier's supply voltage in synchronism with the amplitude envelope of the RF signal. This is known as envelope tracking.
Some types of envelope tracking may include: using direct current (DC) to direct current (DC) converters; power DAC (digital analog converter); and Class “AB” push pull video amplifiers. These are some of the methods used to amplify the amplitude signal. A single amplifier could also be used with Class “A” operation to transfer amplitude information to the carrier envelope. Unfortunately, this is a very inefficient method to transfer envelope energy to the radio frequency (RF) amplifier. Often envelope tracking is used to make slow adjustments to the DC supply only when envelope fluctuations are of relatively low bandwidth. Such an apparatus is not stable or competitively efficient if modulated at a higher rate as needed in present-day cell phones.
Another approach is envelope tracking through adjustment of an amplifier power supply using DC to DC converters. The DC to DC converter output is varied by its output duty cycle in proportion to a desired energy so that the resultant filtered voltage level reproduces an amplitude modulation signal. Unfortunately, a drawback to this approach is that a high modulation rate may not be achieved without distortion and/or stability problems.
In some DC to DC tracking converters the efficiency falls as the load current decreases. This drop is unsatisfactory for optimal modulation restoration techniques since it usually causes performance to fall outside industry specification requirements. Also, another disadvantage to this approach is that such DC converters often require a large, ferrite core inductor to convert the switched energy to envelop power. This undesirably adds to the complexity and cost of the DC converter. Other semiconductor tradeoffs force the issue of reduced efficiency versus power output and bandwidth.
Pre-Distortion
Typically, pre-distortion techniques apply a pre-distorted power amplifier (PA) input signal to a PA. This pre-distorted PA input signal is used to cancel or compensate for inherent distortion of the PA and attempts to improve linearization of the PA. Unfortunately, most digital implementations of pre-distortion utilize digital signal processing (DSP) and software, which can cause resource challenges and consume significant power associated with the management of current PAs, which follow rapid changes in power levels. Moreover, digital implementations of pre-distortion require significant investment of integrated circuit silicon area.
Yet another drawback to pre-distortion techniques is the need to insert a nonlinear module (typically known as a “pre-distorter” module) before the RF power amplifier. This pre-distorter module counters the nonlinear portion of the PA transfer characteristic. Thus the overall system response from input to the output of the PA is linear when compensated by the pre-distortion module. The philosophy of this approach identifies the PA nonlinearity as an undesirable design limitation or weakness which must be removed. Efficiency is not a primary optimization parameter for such schemes.
Adaptive digital pre-distortion is a technique that involves digital implementation of the pre-distorter module and a feedback loop that adapts to changes in the response of the PA due to varying operating conditions. The major drawbacks to this technique are increased power consumption, complexity, size and cost of the system due to the adaptive feedback architecture.
Feedback Loops
As mentioned with respect to pre-distortion above, a feedback loop is a circuit configuration that adapts to changes in the response of the PA due to varying operating conditions. For example, there is a specific type of feedback loop known as a “regenerative feedback loop”.
Typically, any RF (radio frequency, which possesses a rate of oscillation in the range of about 3 kHz to 300 GHz, which corresponds to the frequency of radio waves, and the alternating currents, which carry radio signals) feedback oscillator can be operated as a regenerative receiver if modified to provide a controllable reduction in the feedback loop. It also requires coupling the feedback loop to an incoming signal source, and coupling audio frequencies out of the feedback loop to a subsequent audio amplification stage.
Unfortunately, feedback loops, including regenerative feedback loops, require additional components and therefore, increase the power consumption, complexity, size and cost of the circuit. Also, feedback loops introduce a number of waveform distortions that must be addressed. Thus, the feedback loops can actually introduce additional noise and errors into the system. These unwanted imperfections introduced by the feedback loop result in various waveform contaminations which often offset the benefits.
Polar Modulation
Polar modulation is a modulation technique that uses a modulated signal that is both phase modulated (PM) and amplitude modulated (AM). In one example of polar modulation, the low power modulated signal is split into two components: a phase component; and a magnitude component. The phase and increased magnitude components are then combined using an amplifier.
Unfortunately, polar modulation is an inadequate solution because it requires a relatively large sample rate compared to the signal Nyquist bandwidth and often requires the use of pre-distortion in the phase and magnitude. Feedback loops are often employed further complicating solutions at a significant cost in efficiency.
In addition to the conventional techniques described above, the field of power amplification also includes the use of amplifiers such as: Class “AB” Power Amplifiers; Stage Switching and Doherty Amplifiers; Envelope Elimination and Restoration (EER) Amplifiers; and Outphasing and Linear Amplification with Nonlinear Components (LINC) Amplifiers. Each of these amplification techniques suffers drawbacks that make them unsuitable for use with cell phones.
Class “AB” Power Amplifiers
While Class “AB” Power Amplifiers are a mature and popular technology for high production volume RF amplification circuits, such amplifiers suffer numerous drawbacks. For instance, Class “AB” amplifiers achieve only incremental efficiency gains by adaptive bias control, envelope tracking control, and power supply control. There is a detrimental tradeoff between linearity and efficiency. “Over-the-Air” specifications impose minimum linearity requirements such that precise input power backoff is required to balance linearity and efficiency. (“Input power backoff” is a reduction of the output power when reducing the input power. The efficiency of the power amplifier is reduced due to backoff of the output power, because the amplifier operates in a linear region.) Since input power backoff is waveform dependent, the input power backoff must be increased for higher peak to average waveforms, which reduces efficiency making Class “AB” amplifiers less than ideal for many applications.
Stage Switching Amplifiers and Doherty Amplifiers
Another conventional approach is to use either stage switching amplifiers or Doherty amplifiers.
Stage switching amplifiers are typically implemented with switches or staggered bias control, which can be optimized for efficiency at multiple operating points. Stage switching amplifiers have higher average efficiencies than traditional class “AB” power amplifiers when the output power range traverses the operating points and such amplifiers can also be integrated in various semiconductor processes.
Stage switching amplifiers have a number of undesirable drawbacks. For example, stage switching amplifiers are normally constructed using Class “AB” stages and therefore, have all of the limitations of Class “AB” power amplifiers, some of which were described above. These drawbacks include a tradeoff of linearity versus efficiency and heat dissipation.
Doherty amplifiers are another conventional technique. These amplifiers have increased efficiency for higher peak to average ratio waveforms and the carrier power amplifier PA is biased Class “B” amplification. Typically, with Doherty amplifiers, the carrier PA alone supplies the output power over most of the output power dynamic range. The peaking PA is biased as Class “C” amplification and the peaking PA is “off” during most of the output power dynamic range. The peaking PA and carrier PA of Doherty amplifiers both supply output power during waveform peaks.
Doherty amplifiers suffer numerous undesirable performance drawbacks. For example, they require precise control of the input drive and bias of the carrier and peaking PAs (power amplifiers). They also require precise impedance values to ensure minimum distortion crossover performance as well as having all of the limitations of linear Class “B” power amplifiers. As with the case of stage switching amplifiers, Doherty amplifiers also suffer from linearity versus efficiency tradeoff problems. Additionally, Doherty amplifiers have inadequacies due to input backoff considerations, heat dissipation versus linearity tradeoff.
Thus, both stage switching amplifiers and Doherty amplifiers suffer from numerous drawbacks, some of which have been discussed above. These numerous drawbacks result in less than desired performance for many applications.
Envelope Elimination and Restoration (EER) Amplifiers
EER amplifiers separate the phase and amplitude components from a modulated signal. This type of nonlinear power amplifier technology is employed in the phase signal path, which has no amplitude component. The amplitude signal path has no phase component. EER amplifiers can utilize Class “C”, “D”, “E”, “F” and other nonlinear amplifiers.
EER amplifiers are also referred to as Kahn and/or polar amplifiers and are more efficient than Class “AB” power amplifiers at lower output power levels. The EER amplifier permits the bias and power supply voltages to be controlled so as to optimize power consumption at different power levels. Theses amplifiers can be largely integrated in various semiconductor technologies.
However, EER amplifiers (Kahn and/or polar amplifiers) have numerous undesirable characteristics. For example, ERR amplifiers have extreme difficulty maintaining phase signal path and amplitude signal path alignment. Furthermore, small alignment errors will result in the failure to pass most ACPR/ACLR requirements. Additionally, EER amplifiers generally require feedback to achieve linearity requirements. These feedback mechanisms typically involve polar feedback with separate amplitude correction and phase correction loops or Cartesian feedback loops. As discussed above herein, feedback loops greatly reduce amplifier efficiency. The EER amplifiers which utilize DC to DC converter also require the DC/DC converter bandwidth to be greater than the signal bandwidth and are dependent on input waveform linearity. This is a serious drawback since input waveforms must significantly exceed the output linearity requirements.
Another conventional approach has been to use polar amplifiers with Cartesian feedback. It requires a complex demodulator (I/Q (In-Phase/Quadrature) Receiver) for the feedback path. Furthermore, using this approach can cause errors in the complex demodulator such as Quadrature and Amplitude imbalance that will be present on the output signal. Other drawbacks of this approach include: difficulty maintaining feedback loop stability due to path delays from the baseband to the RF output; the complex demodulator reduces the efficiency; and the requirement that the amplitude envelope reconstruction bandwidth must be much greater than the desired output signal bandwidth.
Outphasing and Linear Amplification with Nonlinear Components (LINC) Amplifiers
Outphasing was first proposed by H. Chireix, (“High Power Outphasing Modulation,” Proc. IRE, Vol. 23, No. 11, November 1935, pp. 1370-1392 as a method of Generating High Power/High Quality AM Signals with vacuum tubes. Starting around 1975, the term “Outphasing” was supplemented with LINC (Linear Amplification with Nonlinear Components) as the technology was adopted for use in microwave applications. Outphasing, or LINC, is a technique that provides In-Phase and Quadrature Phase Baseband Inputs and incorporates transmitter function. It eliminates the traditional RF transmitter to PA (power amplifier) input interface impedance match, filter, and back-off requirements. LINC is able to utilize multiple nonlinear amplifiers in an attempt to increase amplifier efficiency, favorable thermal characteristics and higher available output power. Indeed LINC does not have any amplitude and phase alignment issues that EER architectures do and LINC also has a simple transfer function. Another advantage of LINC techniques is that In-Phase and Quadrature inputs are transformed into two or more constant envelope signal components.
White LINC has some advantages, as discussed above, the technique suffers serious drawbacks. For example, LINC requires power combiner technology with the accompanying large physical size (quarter wave elements are 3.75 cm (1.5 inches) at 2 GHz and 7.5 cm (3.0 inches) at 1 GHz). Secondly, LINC cannot be integrated without large losses, which causes it to be impractical due to semiconductor die size. LINC also suffers from a relatively narrow practical application bandwidth. Moreover, parametric and temperature variations adversely affect performance. LINC has a limited operational temperature range for optimal performance.
Another significant drawback to LINC techniques is a requirement for isolation between branch power amplifiers. While lossless combiners (reactive elements only) have been used, this creates output waveform distortions. Simple Pi-networks have also been used and create undesired output waveform distortions.
Referring back to outphasing, the phase accuracy requirements and physical size are significant drawbacks. For example, at any given power level, to produce quality waveforms, 40 dB of output power dynamic range is desirable. Therefore, two sinusoids with perfect amplitude and phase balance need to vary between 0 degrees phase and 178.86 degrees phase to achieve a 40 dB dynamic power output range. The accuracy required to achieve 40 dB challenges the tolerance of practical circuits in a high volume application. Thus, this technique is not desirable for current cell phone applications.
With respect to the large physical size required by outphasing, as mentioned previously, quarter wave elements are 3.75 cm (˜1.5 inches) at 2 GHz and 7.5 cm (˜3.0 inches) at 1 GHz. With such large size requirements, this approach currently cannot be integrated without large losses, whenever quarter wave combiner techniques are used even on a silicon based substrate. Furthermore, it is impractical due to semiconductor die size. Other drawbacks, similar to those mentioned above include: narrow bandwidth; having real losses that adversely affect efficiency; parametric and temperature variations that adversely affect performance; unit-to-unit performance variations that unexpectedly vary loss, isolation, and center frequency. Additionally, outphasing has a limited temperature range for optimal performance and requires isolation between power amplifiers. Similar to LINC described above, lossless combiners (reactive elements only) have been used and create undesired output waveform distortions. Yet another drawback is that outphasing requires significant branch phase accuracy and branch amplitude accuracy to generate waveforms of acceptable quality.